Protection Electronics And Upgrades for High Voltage Pulse Power Networks

1 Protection Electronics And Upgrades for High Voltage Pulse Power Networks First Semester Report Fall Semester 2014 By Xiang Sun Prepared to partially fulfill the requirements for ECE401 Department of Electrical and Computer Engineering Colorado State University Fort Collins, Colorado 80523 Project Advisor: Dr. Jorge Rocca & Mark Woolston Approved by: Mark Woolston

2 Abstract The function of Pulse forming network (PFN) is store electronic energy in a long period and release these in a short duration. This system had already been developed into three generation and there are 20 units in advanced beam laboratory of Colorado State University. The objective of PFN in the whole research project is to provide high energy pulses to laser amplifiers. This laser amplifier work as the part of research on generating short, coherent extreme ultraviolet pulses. The simmer unit inside of PFN is responsible for creating a stable path through xenon gas which allows an efficient energy transfer from electronic power source to optical energy. This report is aim to describe the process of identifying the problem, make the strategy, verification of our idea, design and creates for the protection module. This protection module should have the ability to reduce the high power which would flow into simmer unit. To achieve this, we came up with two aspects to do that. After the verification, these two aspects need to be combined together. The method taken is to change the diodes considering reverse recovery time and build up a snubber which consists of a RC circuit. These two ways will block and absorb high power respectively. So far, our idea has already been verified and start to the assembling all of the electronic components has started. After finishing this work, I will install the entire module into the PFNs and conduct long-term periodic test to verify the stability and reliability of this protection module.

3 Table of Contents Chapter Subject Page Abstract 2 Table of Contents 3 List of Figures 4 List of Tables 4 Chapter 1 Introduction 5 1.1 - Background 5 1.2 - Motivation 5 1.3 - Goal 5 1.4 - Outline 5 1.5 - Safety Consideration 5 Chapter 2 Conceptual Analysis 7 2.1 - Concept of PFN & Simmer Unit 7 2.2 - Analysis on Problem 8 2.3 - The Strategy for Solving Problem 10 Chapter 3 Practical Implementation 12 3.1 - Software Simulation 12 3.2 - Practical Verification 13 3.3 - Create Protection Module 14 3.4 - Current Development 15 Chapter 4 Conclusion 17 References and Bibliography 18 Appendices Abbreviations, Budgets 19 Timeline 20 Acknowledgements 21

4 List of Figures Figure# Item Page Figure 1 Block diagram of second generation flashlamp pulse power system 7 Figure 2 Simplified plot of a PN diode reverse-recovery period 8 Figure 3 Test graph from PFN 9 Figure 4 Simmer unit: Model 864 9 Figure 5 Damaged simmer unit 10 Figure 6 Proposed Snubber 11 Figure 7 Transient current flow into snubber 11 Figure 8 Simulation for snubber with selected diode 12 Figure 9 Sample protection module 13 Figure 10 Test graph for new results and Original Result 13 Figure 11 Inner space of PFN 14 Figure 12 Designed PCB 14 Figure 13 Assembled PCB 15 Figure 14 Metal bar on solidwork 15 List of Tables Table# Item Page Table 1 Datasheet for diode: Z50SG 10 Table 2 Datasheet for 1N6517& Z50SG 12

5 Chapter 1 Introduction 1.1 Background With the development of various technologies, both industry field and science area are looking for producing accurate nanostructure by finer spatial resolution and finer temporal resolution optical imaging. In Colorado State University, they had already developed a method to generating short and coherent extreme ultraviolet (EUV) pulses by a multiple stage laser system. Basically, this system is need high pulse power system to pump the highest power stage s laser amplifiers. This stage utilizes flashlamp pumped Nd:Glass slab laser amplifiers which need to be drove by 2700 joules of energy with a half-amplitude width of 270 us to four water-cooled pairs on either side of the laser slab. In case of electrical joule losses, there need two pulse forming networks (PFNs) which have the ability to store 2000 joules of energy and releasing 4000A pulses repeatedly, consistently and precisely. [1] 1.2 Motivation With the original design of PFN, once high energy current flow into flashlamp there is no possibility for current can flow into simmer unit since there is three diode to bock that. However, during the reverse recovery time of three diodes, the high energy current sometimes can flow into simmer unit and damage it instantly. With the original operation of 4 first-generation PFNs, the frequency of this happening is too low to attract the designer s attention. After the number of PFN reached to 20 units, this problem occurred normally. For each simmer unit, it will cost about two thousand dollars and the since the frequent damaged on that the total cost could be around 600 dollars. This issue not only results in cost of replacement and maintenance. Moreover, but also caused the suspension on schedule of regular research. 1.3 Goal The project aims to design and create a protection module for simmer unit. This protection module needs to be function well to block the high power flow into simmer unit, and the module itself have be fit into inner of PFN. In order to verify the reliability and stability of this protection module, the long-term periodic test and modification is also required. 1.4 Outline Chapters two introduce the Conceptual Analysis. It includes the theory of PFN and function of simmer unit in PFN. And analyze particular on why simmer unit can be burned by high power and state the proposed strategy. Chapter Three would talk about design and produce technique applied into this project. Chapter Four is regard as conclusion and future work to summarize the lesson learned during this semester.

6 1.5 Safety Consideration *** DANGER *** The voltages, currents and amount of stored charge in this pulse power system are a LETHAL combination. This is not the usual "be careful" statement. THIS EQUIPMENT CAN KILL. Always work in pairs. Preferably with someone who knows CPR. DO NOT TRUST THE FRONT PANEL METER! They frequently fail! Unplug the AC power cords for BOTH the PFN (120VAC) and the General Atomics charging supply (480VAC). ALWAYS MAKE SURE CAPACITANCE IS DISCHARGED before touching ANY of the high voltage areas. Use one of the large high-voltage discharge sticks and have a second person present. You will need to (carefully) remove the front panel to do this properly. If working in the laser lab, it may be acceptable to wire-short-circuit the high voltage at the laser head to ground with a thick, securely connected cable instead. THE SHORT STICK MUST BE APPLIED CONTINUOUSLY DURING ANY WIRE- SHORTING WORK. DO NOT TRUST THE HIGH VOLTAGE DISCHARGE RELAY TO WORK. You MUST discharge the capacitor with the short-stick. Wire the capacitor terminals together (short circuit) to prevent any possibility of charge build up. Note capacitors with dielectric absorption can have charge reappear on them even after being shorted with the stick! Dielectric absorption can get worse as a capacitor ages. SWITCHING PFN POWER OFF DOES NOT MEAN THE UNIT IS SAFE. The General atomics can still charge the PFN capacitor even with the PFN switched off. Be sure to switch off the General Atomics as well. Note the capacitor could still be holding charge! [1]

7 Chapter 2 Conceptual Analysis 2.1 Concept of PFN and Simmer unit The Pulse Forming Network (PFN) is consists of a number of high voltage energy storage capacitors and inductors. And these interconnected components have same function as transmission line. Initially, the capacitor of PFN charged by high voltage DC power supply in a long period. And then release these high power energies with comparatively brief duration. In the advanced beam laboratory of Colorado State University, the research is focus on generating short, coherent extreme ultraviolet pulses which allowed finer spatial resolution and finer resolution optical imaging. This technique can be applied to creating accurate nanostructure. This research system is composed by a series of electronic subsystem. The objective of PFN in this system, is aim to provide high power. The PFN in this system is capable of storing 2000 joules of energy and dumping 4000 A pulses repeatedly, consistently and precisely on time. Its repetition rate is at 4 Hz. [1] In 2006, the first generation of the suitable pulse power systems were designed, produced and developed by Dale Martz. And in 2009, the pulse power systems were developed into second-generation and incorporated with additional capabilities and improved construction techniques. There were four first-generation systems were upgraded with new improvement and six second-generation systems built. With the development of research, the laboratory had built ten more pulse power systems. Simmer Unit Figure 1: Block diagram of second generation flashlamp pulse power system [1]

8 Since this project is mainly focus on the simmer unit, the discussion of other operations of this system is out of the range. Therefore, the following statement is only concentrate on the conceptual idea of simmer unit in PFN. In the figure 1, the simmer units are shown in the yellow box. The objective of simmer unit is aim to establish a stable path through the gas to allow the transfer of energy from electrical power source to optical energy repeatedly. 2.2 Analysis on problem In the original design, there are three series diode are connect to the output terminal of simmer unit. Its reverse direction when the high power approaches to these diodes. Therefore, the high energy power from high voltage capacitor will be only flow into flashlamp. This trace had shown in the figure 1 in the form of blue arrow line. Figure 2: Simplified plot of a PN diode reverse-recovery period [2] However, as figure 2 shown, the diode has a specific characteristic which called reverse recovery effect. It s the switching loss which depends on the reverse-recovery time (t r ). This effect is following the forward conduction in a p-n type diode, and a reverse current can realize the conduction through diode in a short period. During this process, the diode does not have the ability to block the current until the mobile charge in the junction is depleted [3]. This effect can be ignored only when the slew rate of the current is not very severe. [3] The reverse recovery effect for this circuit also cannot be neglected. Since this reason, during the reverse recovery time of three series diode, the high voltage energy has the ability to access the simmer unit. In the figure 1, there two green arrow line that shunted from the blue arrow line. It illustrates the fact that the high power not only flow into flashlamp, but also goes into simmer unit through three series diodes.

9 Figure 3: Test graph from PFN The figure 3 obtained by practical test on PFN. The electronic spike shown in the graph is describing the high power approaching to simmer unit. The peak of the electronic spike is approach to 0.5 6 1000(amplification coefficient of high voltage probe) = 3000 V Max. Open Circuit Voltage Figure 4: Simmer unit: Model 864 As the figure 4 shown, the maximum open circuit voltage is 1500 V. The peak of electronic spike has been twice of this value. This electronic spike has direct damage on simmer unit, the diode of simmer unit can be blew up. By comparison on destroyed diode and normal diode, it s clearly to show the damage to simmer unit. There is fact that destruction of simmer unit not happened very frequently. The reason for this happened is its will result in slight damage on the simmer unit once the electronic spike flows into to

10 simmer unit. So, it is accumulated damage on the simmer unit other than instant vital destruction to simmer unit. Destroyed Diode Normal Diode Figure 5: Damaged simmer unit With the development of research, there are 20 PFNs in total and five to six simmer units could be break down in each year. Once a simmer unit broke down, the replacement of a new simmer unit is 600 dollars. The diagnostic and maintenance for destroyed simmer unit affect the regular research schedule. 2.3 The strategy for problem solving Basically, there are two aspects to resolve this problem: First, it s selecting different diode to shrink the reverse recovery time. Based on analysis and test, the reverse recovery time has direct effects on this problem. By shrink the reverse recovery time, its aim to reduce electronic spike to flow into simmer unit. Originally, the used diode is Z50SG which produced by Voltage Multiplier Inc. its reverse recovery time is 3000 ns (Figure 5). Table 1: Datasheet for diode: Z50SG [4]

11 Another proposed method is applying a snubber to restrain the high power flow into simmer unit. The snubber here is consisting of a RC circuit with a diode. The location of snubber circuit is shown in the table 1 (Figure 6). Figure 6: Proposed Snubber In this RC snubber circuit, is aim to limit dv. High voltage here indicates high energy dt is present. Therefore, once the high voltage approaching to simmer unit, the voltage across the three series diode has been raised up (Figure 8) resluting in transient current which flows into snubber (Figure 7). As the function V = C di shown, the capacitor is dt going absorb the electronic energy. After this energy discharged to the resistor, this electronic energy will be transformed into thermal energy. Figure 7: Transient current flow into snubber

12 Chapter 3- Practical Implementation 3.1 Software simulation LTspice was used to realize the following objectives: Simulate preexisting system to get the original performance of PFN Examine our two strategies on this problem Find the proper components: diode, resistor and capacitor. Input Voltage 5000 V Output terminal of Simmer Unit Figure 8: Simulation for snubber with selected diode The figure 8 shown is the schematic for snubber with replaced diodes and simulation graph for this protection module. As the arrow pointed, the input voltage is 5000 V and there is only 1500 V at the output terminal of simmer unit. The rest of energy is reduced by 3 series diode (Total Working Reverse Voltage: 3*5000=15000 V) and RC snubber. The strategy was verified. For the proper components, the values decided by repeat simulation. Resistor: 10M Ω Capacitor: 1000p F Diode: 1N6517 (From Voltage Multiplier Inc.) Table 2: Datasheet for 1N6517& Z50SG

13 The reverse recovery time for 1N6517 is 70 ns. (Table 2). The table 2 indicates that these two diodes have similar values for the most factors and the biggest difference is reverse recovery time. 3.2 Pratical Verification Use the sample components to verify the concepts on PFN. Figure 9: Sample protection module Plastic shield is used to avoid the potential damage from short to the high voltage capacitor. Test tool: high voltage probe & digital storage oscilloscope. Test Results: Figure 10: Test graph for new results and Original Result In figure 10, the new result has big differences with original result. In the figure 10, the electronic spike had been cut off pretty much, the period of this spike had also been shortened. This result indicates the proposed two strategies are succeeded.

14 3.3 Create Protection Module The inner space of PFN is very limited and complex. In order to make this module fit into PFN (Figure 11), it s better that the protection module lay out onto PCB. Figure 11: Inner space of PFN The following aspects are important for the design of a PCB: a. the size of each component; b. the distance of each component; c. using separate connectors; d. using two layers; e. use an high voltage wire to jump over the traces on the board. All these aspects are considering the risk of high power. Figure 12: Designed PCB The figure 12 is the designed PCB schematic.

15 3.4 Current Work For the current development, I m working on assemble all components onto 20 PCBs (Figure 13). Since the limit of space and mistake on size of PCB, there need some modifications for PCB. Besides, I am also using solidwork to design a metal bar to attach PCB to PFN (Figure 14). This metal is aim to using the screw hole to fix the metal onto itself and the metal also use screw to fasten with PFN. Meanwhile, it also realizes the function of connection between PCB and ground. Figure 13: Assembled PCB Figure 14: Metal bar on solidwork

16 Chapter 4-Conclusion Challenge: This is a practical project. Therefore, during my design and installation I need to take every aspect into considerations before I take action, because some mistake will make me into a bigger trouble and waste on time or money or both. And there are 20 PFNs, each have some differences since all these PFNs not produce at the same time. What s more, there also exist high power risk, so each time I m trying to open the box there have one more people to accompany with me. Learned: In this project, I also learned much software, such as LTspice, ExpressPCB, Solidwork. I learned some professional skills as an electrical engineering. Besides, I realize when I am doing regular project work, I not only need to know how to conduct the project, but also I need to basic engineering theory. Future work: For the next semester, the things I need to do are as list: Assemble all the components onto 20 PCB Use a standard to unify the part of inner space of PFN Design and produce a metal bar and plastic shield Install all 20 assembled PCBs, metal bar, and plastic shield into PFN Do the original test for each of these protection modules Make a long-term, periodic test to verify the stability and reliability of 20 protection modules. Design a remote control module to realize some basic operation for PFN. For a part, since we are going to use a different connector for PCB, I need to drill the hole by myself for total 20 PCB board. There are 15 components for each PCB. For b part, the reason to do so is because not all the PFN were produced at the same time. There are some high voltage wire blocked the space where I need to locate PCB and I need to move all them. When I am going to unify this space, I also need to be care of some high voltage risk. For c part, this metal bar is used to attach PCB to PFN and make PCB connect to ground. This metal itself should also fit into PFN very well. In terms of plastic shield, it should have the ability to protect protection module from potential damage from short to the high voltage capacitor. And this plastic shield needs to be connected by plastic shield and fit into PFN well. For d part, just need to be careful on installation. For e part, the test is aim to verify the connection and functionality of these PFN. If there are some problems, I could take some modifications as soon as possible. For f part, this is a long time work. I need to take every data very detail, and the test would be once a week for the whole PFN. I also need to notice the each of them. I am not only to check the functionality of these PFN, but also going to record the change of shape.

For g part, this process is aim to realize some basic function of PFN, such as remote control on turning on and turning off. Because the laboratory and the PFN are not located at same room, it s inconvenient for the researcher to launch the equipment. 17

18 References [1] Mark Woolston, Power Electronics for High power Lasers Pulse Power System for a Flashlamp Pumped Laser Amplifer, Senior Design Project Report, Colorado State University, Fort Collins (2006) [2] Maxim Integrated Products, Inc. An Efficiency Primer for Switch-Mode, DC-DC Converter Power Supplies. Dec 23, 2008. http://www.maximintegrated.com/en/app-notes/index.mvp/id/4266 [3] ECEN5817. Diode reverse recovery in a boost converter. ecee.colorado.edu [4]Voltage Multiplier Inc. Z25SG-Z100SG Datasheet. http://www.voltagemultipliers.com/pdf/z25sg-z100sg.pdf [5] Voltage Multiplier Inc. 1N6513_1N6515_1N6517_1N6519. http://www.voltagemultipliers.com/pdf/1n6513_1n6515_1n6517_1n6519.pdf Bibliography N/A

19 Appendices Appendix A Abbreviations Cap PFN Tr Capacitor Pulse Forming Network Reverse Recovery Time Appendix B Budget All expenditures are paid by NSF grants. $ 25 spent by Senior Design funds Item Quantities Unit Price ($) Total Price ($) Diode 8*20+10 13.65 2320.5 Resistor 2*20+10 0.26 13 Capacitor 2*20+10 2.8336 141.68 PCB 1*20+2 -- 734.5 Controller 3*20+40 -- 115.1 Metal Bar 20 4.33 86.6 Plastic Shield 20 5.06 101.2 Total = $ 3512.58

20 Appendix C Timeline Time Period (#week) Objectives (Completed if Underlined ) Beginning of the semester to September 14 th Take basic research on the project From September 15 th to 21 st Do the test on the built circuit, collect corresponding data and analysis data Comparing current diode with ideal diode From September 22 nd to September 28 th Plug protection circuit into PFN and record corresponding data Replace the diode with the better version that we have and test it, record data Order the better diode for the whole system From September 29 th to October 12 th Simulate the protection circuit on the software based on recorded data and adjust it, then apply these adjustments into the circuit to do the test and record results. Repeat this procedure over and over again until the circuit reach perfect situation From October 13 th to November 2 nd Design PCB board as the module for both protection circuit and diode and order the PCB board From November 3 rd to December 12 th Assemble all diode, resistors, capacitors and connectors onto PCB board, and install all these units into PFN circuit and test it, debug it From December 13 th to the end of semester Be prepared for the next semester 2015 spring semester Start designing a remote control system on PFN and periodic test

21 Acknowledgements I would like to thank my supervisor, Dr. Jorge Rocca for the opportunity he given to me and the encouragement. I would like to thank my project advisor Mark Woolston for his support, patient and many professional skills to help me develop my project. And I would like to thank Alex RockWood, Yong Wang, Shoujun Zhao and Liang Yin. They gave me enough support on regular work.